MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

A good day to everyone.

Consider the following "Conjecture":

If $a, b \in M \subset \mathbb{N}$, then $1 < a < b$ and ... [plus some more conditions on $a, b$ and $M$...] if and only if $\sigma(a) < \sigma(b)$.

I have two questions at this point:

(1) What properties should elements of the set $M$ have to satisfy this "Conjecture"?

(2) What happens to the sets in $M$ if you restrict to the case $\gcd(a, b) = 1$? (Of course, the sets in (1) should have greater asymptotic density than the sets in (2), but can this notion be made more precise?)

Now, for the motivation: An example of a set $M$ satisfying this "Conjecture" is any pair $(A, B)$ of relatively prime factors of an odd perfect number given in the Eulerian form $N = {q^k}{n^2}$ (modulo some exceptions), where $\gcd(q, n) = 1$ and $q \equiv k \equiv 1 \pmod4$.

Let $$I(x) = \displaystyle\frac{\sigma(x)}{x}$$ be the abundancy index of $x \in \mathbb{N}$. Then by enumerating all possible permutations of the set

$$q^k, n, \sigma(q^k), \sigma(n) \in \mathbb{N}$$

we know that exactly one of the following holds (because of $1 < I(q^k) < I(n)$):

$$[1] \hspace{0.2in} q^k < n < \sigma(q^k) < \sigma(n)$$ $$[2] \hspace{0.2in} q^k < \sigma(q^k) < n < \sigma(n)$$ $$[3] \hspace{0.2in} n < q^k < \sigma(q^k) < \sigma(n)$$ $$[4] \hspace{0.2in} n < q^k < \sigma(n) < \sigma(q^k)$$ $$[5] \hspace{0.2in} n < \sigma(n) < q^k < \sigma(q^k)$$

Last question: Which of these five "configurations" could we eliminate?

share|cite|improve this question
Arnie, where did you pick up the notation I(n) for the abundancy index? I haven't seen that used before. – Charles Mar 7 '11 at 0:10
@Charles, as far as I know Richard Ryan and Judy Holdener uses that notation in their papers. For a (comprehensive?) list of resources on the abundancy index, you can refer to Walter Nissen's page at – user11235813 Mar 7 '11 at 1:35
Thanks! I've read most of the papers linked there, including Holdner's "A theorem of Touchard on the form of odd perfect numbers", but not the Ryan paper. – Charles Mar 7 '11 at 2:45
You're most welcome =) – user11235813 Mar 7 '11 at 3:40
up vote 3 down vote accepted

I think this is a ludicrously broad question. For (1), you could take $M$ to be the set of "champions" for the divisor function, $M=\lbrace1,2,3,4,6,8,10,12,16,18,20,24,\dots\rbrace$, which is - no conditions on $a$, $b$ necessary. At the other extreme, you could take $M$ to be the whole of the positive integers, with the condition $a\mid b$. There may be some interesting answers in between, but this strikes me as more of a fishing expedition than a serious question.

share|cite|improve this answer
Thanks Gerry! Rest assured that it IS a serious question. =) – user11235813 Mar 5 '11 at 23:53
+1 for the first and last sentences. – Yemon Choi Mar 6 '11 at 3:39
Thanks Yemon! @grin@ – user11235813 Jun 12 '11 at 21:38

Let $N = {q^k}{n^2}$ be an odd perfect number (i.e. OPN) given in the so-called Eulerian form (i.e. $q \equiv k \equiv 1 \pmod 4$ and $\gcd(q, n) = 1$).

Note that $k = 1$ rules out [1] and [4], because $q$ and $\sigma(q) = q + 1$ are consecutive integers. I am guessing a possible approach will be to derive [a] condition(s) equivalent to $k = 1$ and work from there.

A sufficient condition for $k = 1$ is $n < q$ (or $n < q^2$). I will be most surprised if somebody would be able to establish that either of these two is also necessary for $k = 1$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.